Sequence Length Distribution in Segmented Block Copolymers

Sequence Length Distribution in Segmented Block Copolymers. L. H. Peebles Jr. Macromolecules , 1974, 7 (6), pp 872–882. DOI: 10.1021/ma60042a034...
0 downloads 0 Views 1MB Size
872 Peebles

Macromolecules

domains of different densities. This conflict does not in fact occur; these domains can exist as long as they do not interfere with the statistical properties of each molecular chain. All the models of polymeric systems now proposed must take these facts into account.

Acknowledgments. This work was made possible by the CNRS who gave financial aid in the form of an ATP and by the allocation of beam time on the D 11 apparatus a t the Institute Laue-Langevin (Grenoble, France). We are grateful to Dr. K. Ibel of this Institute for discussions about the use of the apparatus and to M. Rebesco and Mr. Barthelemy for assistance in running it. We would like to thank Dr. D. Cribier (CEA-SRM) for his constant help and fruitful discussions during the experiments. We also appreciate the help of Mr. M. Herbert (CEA Ddpartement des Radio616ments) in obtaining deuterated monomers. References and Notes (1) (a) Service de Physique d u Solide et de Rbsonance MagnBtique; (b)

Centre de Recherches sur les Macromolbcules; (c) Institut Max Von Laue-Paul Langevin; (d) Laboratoire de la Matiere CondensBe; (e) Service de Physique ThBorique. (2) (a) W. Kuhn, Kolloid-Z., 68, 2 (1934); 76, 258 (1936); (b) H. M. James, J . Chem. Phys., 15, 651 (1947); H. M. James and E. Guth, 11, 455 (1943); 15,669 (1947).

(3) P. J. Flory and J. Rehner, J.Chem. Phys., 11,512 (1943). (4) J. D. Ferry, “Viscoelastic Properties of Polymers,” Wiley, New York, N. Y., 1970. (5) W. Pechold, Kolloid-2. 2. Polym., 228,l (1968). (6) G. S. Yeh, J.M Q C ~ OSci., ~ O Phys. ~ . 6,451 (1972). (7) R. G. Kirste, W. A. Kruse and J. Schelten, Makromol. Chem., 162, 299 (1973);Polymer, t o be submitted for publication. (8) D. G. Ballard, J. Schelten, and G. D. Wignall, Eur. Polym. J . , 9, 965 (1973). (9) V. F. Turchin, “Slow Neutrons,” Israel Progress in Scientific Transactions Ltd., 1965. (10) I. I. Gurevitch and L. V. Tarasov, “Low Energy Neutron Physics,” North-Holland Publishing Co., Amsterdam, 1968:~ (11) Tables of MIT, 1972, collected by G. Sholl. (12) J. P. Cotton, Ph.D., University Paris VI, 1973; CEA, No. 1763. (13) P. Ageron, J. M. Astruc, H. Geipel, J. Verdier, CEA (France) BIST, 166.17 (1972). (14) D. Cribier, B. Jacrot, A. Lacaze, and P. Rousseau, IAEA Vienne, 1960. (15) B. Farnoux, B. Hennion, and J. Fagot, IAEA Vienne, 1968. (16) P. Ageron, P. A. Blum, CEA (France) BIST, 166,35 (1972). (17) G. Gobert, CEA R, 2981 (1966). (18) G. Degenkolbe and H. B. Greiff, Kerntechnik, 15,437 (1973). (19) A. Guinier and G. Fournet, “Small Angle Scattering,” Wiley, New York, N. Y., 1955. (20) M. Szwarc, Makromol. Chem., 35,132 (1960). (21) D. J. Worsfold and S. Bywater, Can. J. Chem., 38, 1891 (1960). (22) A. L. Gatzke, J . Polym. Sci., Part A-1, 7,2281 (1969). (23) Z. Grubisic-Gallot, M. Picot, Ph. Gramain, and H. Benoit, J. Appl. Polym. Sci., 16,2931 (1972). (24) J. des Cloiseaux, J. Phys., 31,716 (1970). (25) P. Debye,J. Phys. Colloid Chem., 51,18 (1947). (26) C, Strazielle, et al., in preparation. (27) C. Loucheux, G. Weill, and H. Benoit, J . Chimie Phys., 540 (1958).

Sequence Length Distribution in Segmented Block Copolymers L. H. Peebles, Jr.’ Office of Naval Research, Boston, Massachusetts 02210. Received March 18,1974

ABSTRACT: The effect of the nonequivalent reactivity of the two functional groups in a diisocyanate on the sequence length distribution in segmented block copolymers is calculated on a kinetic basis for a number of polymerization recipes. In the two-stage recipe, excess diisocyanate monomer first reacts completely with an intermediate molecular weight prepolymer then a low molecular weight extender is added. Alteration of the definition of the hard block segment in the two-stage process demonstrates that the distribution follows the most probable form. The heterogeneity index is less than 2 because of the relatively short sequence lengths. The number average length of the hard block segment under stoichiometric conditions a t 100% conversion is equal to A 1/B 1 when the isocyanate groups are equally reactive, where A 1 is the initial concentration of diisocyanate and B 1 is the initial concentration of prepolymer. When the reactivity of the first isocyanate group becomes appreciably greater than that of the second isocyanate group, the number average length of the hard block segment approaches ( A1 - B 1 ) / B1. The sequence length distribution of the hard blocks, the amount of internal sequences, and the molecular weight of the polymer can be varied by alteration of the polymerization recipes and the method of mixing the components.

Segmented polyurethanes are usually prepared by reacting a diisocyanate (AA) with a hydroxy-terminated polyester or polyether (BB) to form a soft block segment. The initial content of diisocyanate (AA) exceeds that of the prepolymer (BB) so that there will always be an excess of unreacted diisocyanate. The resultant polymer is then extended with a low molecular weight diol or a low molecular weight diamine (CC) to form a hard block segment. During the fabrication of the final article, the hard block portions segregate from the soft block segments to form a microheterogeneous composite.2 The reactions involved in this process are rather complex. The urethane bonds I formed from the reaction of isocyanate with hydroxyl and the urea bonds I1 formed from the reaction of isocyanate with amine are somefiat reversible. If any water is present in the system, it will react with an isocyanate group to eventually form an amine plus COS gas; the amine can then react further with an isocyanate group. Branching and cross-linking reactions can occur by

reaction of an active hydrogen from a urethane or a urea group with an isocyanate group to form allophonate I11 or biuret IV groups, respectively. In the presence of catalysts, the isocyanate groups self-polymerize to uretidione V, isocyanurate VI, or a linear polymer VI1 of high molecular weight. Furthermore, it has been demonstrated that the reactivity of a given isocyanate group in a diisocyanate monomer may depend strongly on whether or not the other isocyanate group has reacted.3a Because of these complex reactions, little attention has been paid to the kinetics of polymerization and, in particular, to the precise composition of the soft and hard block segments. The mechanical properties of segmented polyurethanes depend strongly on the micromorphology of the finished polymer which in turn depends upon the hard and soft block composition and on the processing details. A very precise study of the effects of segment length and distribution has been undertaken by Harrell,3b who combined monodisperse hard segments of various lengths with soft

Sequence Length Distribution in Segmented Block Copolymers 873

Vol. 7, No. 6,,Vovember-December 1974 0

0

0

II

I1

R-NHCOR’

It-SHCNHR’

I

0

R’

II I

I1

RNHC-N-C--OR”

I1

Table I .Relative Reactivities for the Consecutive Reactions of Diisocyanates with Two Representative Alcohols

I11

Relative rate constants

0

Ethanol4

I1

R-N

0

12’

Diisocyanate

\c,

N-R

‘c/

0

11 I II

RNH---C-l\I-NHR”

11

0

E

V

VI

VI1 H

VI11

O

IX

I N C 0 XI1 = TDI

ki

rn -Phenylene p-Phenylene 4,4‘-Diphenylmethane 2,6 -Tolylene 2,4-Tolylene 3,3 ’-Dimethyl-4,4 ’diphenylmethane Dur ene 1,5-Naphthalene 1,3-Xylylene 1,4-Xylylene 5 -tert-Butyl-1,3 xylylene lf6-Hexamethy1ene Phenyl isocyanate (standard)

l-Butanolb

k,h*

2.3

3.2

ki

k i h

11 7.7

8.4

2.4

2.2

3.0

25

4.9

2.7 0.20

3.5 2.5

0.18

1.9

0.19

2.7

9.2 2.9 6.1 11.9

0.41 0.05

2.4 2.6

0.001

2.0

(1.o) (1.o) In toluene a t 30”, excess ethanol.6 In toluene a t 40”, with triethylamine catalyst.’

average molecular weight of the polymer before and after the addition of extender, the average length of the hard block segment in the extended polymer, and the distribution of the hard block segment length. The variables to be considered are the initial concentration of the reactants A 1, B1, and C1; the relative reactivity of the diisocyanate groups, w ; the conversion of the extension reaction; and the sequence of combining the components AA, BB, and CC. General T r e a t m e n t

segments under stoichiometric conditions. Harrell found that thermal and mechanical properties of the polymers did depend on the length and distribution of both the hard and soft segments. An alternative method is to vary the hard block composition by using different polymerization recipes. Several recent publications have been addressed to this problem, but the hard and soft segments are specified by the initial composition of the reactants and not by the statistical probability based on the reactivity of the functional groups i n ~ o l v e d . ~ Let us consider fo:r the moment the reaction of 2 equiv of AA with 1 equiv of BB and assume that all molecules are equally reactive independent of molecular weight. The statistics of the system require that when all of BB has reacted, all molecules are capped with A-type ends. The mole fraction of unreacted AA monomer is 0.5, the mole fraction of AABBAA is (0.5):!, that of AA(BBAA)* is (0.5)3,etc. It is easily seen that if an A-type end of a monomer reacts faster than an A-type end of a polymer molecule, then molecules of BB will tend to be capped by AA, leading to a decrease in the amount of high molecular weight polymer and in the amount of unreactedl AA. On the other hand, if the A-type end of a monomer reacts slower than an A-type end of a polymer molecule, the distribution is shifted toward an increase in the amount of high molecular weight polymer and in the amount of unreacted AA. Thus, the composition of the hard block segment will depend on the relative reactivity of the isocyanate units, the initial concentration of the reactants, and the procedure used to mix the reacting components. It is the purpose of this paper to calculate the number

Table I, taken from Lenz,j gives the ratio of the first reaction rate constant of an isocyanate group to the second reaction rate constant for a number of diisocyanates with two aliphatic alcohols. One can speculate that the electron donating or withdrawing character of a substituent on a benzene ring can materially influence the reactivity of a given isocyanate; for instance, p - phenylene diisocyanate VI11 as compared to 4-(imino carbony1oxyethane)phenyl isocyanate IX. On the other hand, it is not easy to see how the reactivity of one end of hexamethylene isocyanate can be influenced by the structure a t the other end by invoking electronegativity effects. A proximity effect is suggested by 0

/NcO

R

\ NCO

+

HOR’

k’

1I

R

\

X

NCO XI

0

0

It

R

/”CR

\

NCO

+

HOR’

k

II

R

/“CR‘

\

(2)

NHC R’

It

0

eq 1 and 2, where the proximity effect is rate controlling5 but no spatial details are given. Another way of expressing eq 1 and 2 is to state that the

871 Peebles

Macromolecules

reactivity of the monomer X is different from that of the polymer XI. The relative reactivity of the diisocyanate groups is p = k’lk. The above arguments can be applied to any symmetrical molecule such as those cited in Table I. This will be called the symmetrical case. However, the situation might be different in the case of the nonsymmetrical molecules such as tolylene 2,4-diisocyanate (TDI), XII. The reactivity of the two isocyanates in the monomer could differ considerably owing to steric effects. In addition to the steric effect, a proximity effect as indicated in eq 1 and 2 can further affect the reactivity. If such a situation exists, there must be four reactivity constants. The latter case is too difficult to be considered here, hence we shall be concerned only with a steric effect for the nonsymmetrical case. In both cases, one has to be concerned with the definition of a hard block segment. In the simplified reaction scheme to be considered, only urethane groups are formed. Every prepolymer diol unit ends in a urethane group because the initial reaction between excess AA and BB is assumed to proceed until all B-type ends have reacted, a t which point X1O unreacted AA units remain. The extension reaction may be conducted with an excess of either AA or CC and may not go to complete reaction of available groups. Hence there may exist unreacted A- or C-type ends. The reacted A- or C-type ends also form only urethane groups. With these restrictions the definition of a hard block segment can be considered. When a diisocyanate unit is sandwiched between two prepolymer units -AA-

-

type 1

where represents a flexible prepolymer BB unit, it probably will not be involved in phase segregation, hence is considered part of the soft block segment.3b Similarly, the external segment type

-AA type 2

is considered as part of the soft block segment, hence all units of the form AA(BBAA),-1, where n > 1, are considered as soft segments. However, one can argue whether the internal segment type -AACCAA-

type 3 would be crystallizable; the choice may depend on the particular combination of diisocyanate and extender molecules chosen to form the segmented polyurethane. The next possible internal segment would be of the form -AACCX,CCAA

-

type 4, s = 1 where X 1 represents reacted diisocyanates that remained as unreacted monomer at the end of the AA-BB reaction, while AA represents diisocyanates that reacted at one end with a prepolymer unit during the AA-BB reaction. Larger internal segments can also be formed -AAtCCX,),CCAAt>pe 1

If segment type 3 is considered as part of the soft block, then there are exactly s XI units in the hard block segment type 4. This number is called the “counted unit” to avoid confusion with the usual term “repeat unit” because a variety of ways of counting the monomers in a hard block seg-

Table I1 Types of Segments Present in Segmented Polyurethanes, and Definition of Repeat Units i n Hard Block Segments No. of counted units i n a

hard block segment Model Segment types

(1) -AA--“ (2) -AA (3) -AACCAA(4)-AA( CCXi)&CAA(5) -AA(CCX,),

IIC

Ib

IIId

0 0 0

0 0

0 0

1

S

s + l

1 s + l

S

S

S

S s + l s + l (6) -AA(CCX,),CC s + l s + l s (7) X*(CCX,), S S S (8) (CCXi), (9) (CCx,),CC S S s + l represents a prepolymer segment. The unit that is being counted is a diisocyanate residue XI derived from unreacted monomer remaining after the AA-BB reaction. The counted unit is XI plus one isocyanate unit from each soft block segment unless it is segment type 5. d The counted unit is an extender molecule.

-

ment will be used. If segment type 3 is considered part of the hard block segment, then there are s 1 reacted diisocyanates and s + 1 reacted extender molecules in segment type 4. Table I1 gives a summary of the segment types so far discussed and five more types that are also hard block segments. Model I is the designation when segment type 3 is considered soft and only the XI units are counted. Model I1 is the situation when segment type 3 is considered hard and the number of diisocyanate units are counted. In both of these models, unreacted diisocyanate monomer is considered as a hard block portion in order to simplify the mathematics. Model I11 is the situation when the number of extender units are counted in a hard segment. The only difference between models I1 and I11 is which unreacted monomer is counted as part of the hard block segments. At high conversion, under stoichiometric conditions, the concentration of unreacted monomer approaches zero, and the difference between models I1 and I11 vanishes. The importance of defining so many models will become apparent in the discussion. With these definitions of the symmetrical and nonsymmetrical cases and the various models of the hard block segment, we may proceed to consider the amount of unreacted diisocyanate at the end of the AA-BB reaction and the distribution of repeat units in a hard block segment under the following assumptions: (1) the reactivity of one end of a symmetrical diisocyanate monomer depends only on whether or not the other end has reacted; ( 2 ) the reaction between a diisocyanate and a diol is irreversible; (3) if one of the diisocyanate ends has reacted, whether the monomer is symmetrical or not, the reactivity of the unreacted end is independent of the molecular weight of the polymer to which it is attached; (4)the extender is a low molecular weight diol; (5) the reactivity of any hydroxyl group is independent of molecular weight; and (6) no other reactions occur. Because the derivations require great attention to detail, which is not required to follow the discussion, the results and their implications will be presented now. The derivations are in the appendices.

+

Results a n d Discussion Equations A34, A48, A51, A55, A58, A61, and A64 for the symmetrical case, model I, and B9-Bll for the nonsymmetrical case, model I, were solved on an IBM 1130 computer by the RKGS subroutine given in the Scientific Subroutine

Sequence Length Distribution in Segmented Block Copolymers 875

Vol. 7, No. 6, .ouenaber-December 1974

Table IV Values of X10,Unreacted Monomer at the End of the AA-BB Reaction for the Nonsymmetrical Case

Table I11 Values of X l 0 , Umreacted Monomer at the End of t h e AA-BB Reaction for the Symmetrical Case

Initial concentration of AIb

Reactivity ratio"

P= 2

k'/l? $?'

1 3 12 36

0.713 0.500 0.306 0.139 0.066

Initial Concentration of AIb 3 4 5 1.565 1.333 1.159 1.049

2.471

2.250 2.108 2.031 2.011 2.000

3.404 3.200 3.081 3.022 3.008 3.000

6

4.356 4.167 4.065 4.018 4.006 4.000

1.017 1.000 a T h e rate of the polymer reaction, k , differs from that of the monomer, k'. h B1 = I mol/l., C1 = A1 - B1. These are defined as stoichiometric conditions.

=

o

2

3

4

5

6

3 12 36

0.434 0.251 0.132

m

O

1.274 1.129 1.053 1.000

2.200 2.088 2.035 2.000

3.158 3.067 3.026 3.000

4.130 4.054 4.021 4.000

ILa

Package (see paragraph at end of paper regarding supplementary material).'l The program also computed values of conversion; unreacited diisocyanate monomer; unreacted extender; the number average number of diisocynaate molecules in a hard block segment, s; the number average molecular wieght, M T ; and the number of monomer units per ~ . solutions can be obtained polymer molecule, r i ~ Analytical for models I1 and I1 [. Tables I11 and IV give the values of X1O for a variety of p and A1 values and uith B1 = 1 mol/l. and C1 = A1 - B1. For the symmetrical cae,e, X1O decreases with increasing p , approaching the limit A1 - 2B1 as p approaches infinity or as A1 increases. If the value of p is less than unity (that is, polymeric A-type ends react faster than the monomeric Atype ends) the result of the AA-BB reaction is a high molecular weight oligomer with a large excess of unreacted monomer, as predic1;ed in the introduction. For the nonsymmetrical case (Table I\') values of Xl0 for p = lly are equivalent to those for p = 3 because it makes no difference if the Ao end of a diisocyanate monomer of the form A n A is three times faster than the A end or if it is three times slower than the A' end. Values of Xl0 for p = 1 are identical with those given in Table 111. The X10 values for the nonsymmetrical case do not decrease as rapidly with increasing p a:; in the symmetrical case but still approach the limit A1 .- 2B1 at high values of p or AI. The number average molecular weight for the extended polymer is computed by assuming molecular weight values of 250 for AA, equivalent to MDI, OCN-Ph-CHZ-Ph-NCO; 2000 for BB, a common value of commercial material for either a polyester or a polyether; and 90 for CC, equivalent to butanediol. Table V gives the values for the symmetrical case a t 0, 95, and 99% conversion. The initial values are consistent with the concentration of unreacted monomer; high molecular weight oligomers result in more unreacted monomer. As the ratio of Al/B1 is increased, the molecular weight drops, owing to the increased amount of monomer to compete with the partially formed oligomer. At high values of p , the oligomer is primarily composed of one unit of prepolymer capped with two units of diisocyanate monomer. At 95 and 99% conversion, there is little change in i%f~ or l f z with ~ increasing p. The primary difference among = l/3, 1, and the rest is due to small amounts of unreacted diisocyanate remaining in the mixture. There is essentially no unreacted diisocyanate a t higher values of p. The decrease in M,r and m~ with increasing AI is due in part to the increased molar concentration of the low molecular weight monomers relative to the high molecular weight prepolym-

The initial rate constants of monomer are different. h Stoichiometric conditions. Q

er and in part to a lower value of m~ at the start of the extending reaction. Table VI presents the number average molecular weight and degree of polymerization of oligomer at 0% conversion for the nonsymmetrical case. The differences between the entries in Table VI us. Table V are due to the differences in Xl0. Values of the parameters a t 95 and 99% conversion should agree closely with those in Table V because the values of unreacted diisocyanate and extender are almost identical. The number average number of diisocyanate molecules in the hard block segments, S , is given in Table VI1 for the symmetrical case, model I, and in Table VI11 for the nonsymmetrical case, model I, both a t 95,99, and 100% conversion. As conversion increases, increases, not because of the depletion of unreacted monomer but because of the coupling of preformed sequences to make longer sequences. The nonsymmetrical case gives slightly higher values for 3 , owing to the higher values of XIO. Both cases tend toward an s of (A1 - B1)/B1 as p becomes large, b u t s equals A1/Bl when p = 1 at 100% conversion. Several authors have assumed that s is equal to A1 - B1 when B1 = 1, based on a simple stoichiometric approach. This value is called the "ideal" case in Tables VI1 and VIII. In models I1 and I11 for both the symmetrical and nonsymmetrical cases, s under stoichiometric conditions at 100%conversion gives

Substitution of X1O obtained from Tables I11 and IV into eq 3, recalling that C1 = A1 - B1, yields exactly the values s given in Tables VI1 and VI11 at 100% conversion! This remarkable outcome must be due to the distribution of the monomers that make up the hard block segment. Reference to Table I1 shows that model I counts the XI units, which, except for segment types 5 , 7, and 8, are exactly the number of CC units less one. At 100% conversion, only one molecule of infinite molecular weight remains which contains only segment type 4 groups. The number average number of extender molecules per hard block segment is independent of whether one counts all of the extender molecules per segment or all of the extender molecules less one per segment. Let G ( s ) be the concentration of all extender molecules per hard block segment and

also

Macromolecules

876 Peebles

Table V Number Average Molecular Weight and Degree of Polymerization of "Polymer" as a Function of the Extending Reaction for the Symmetrical Case (see eq A23-26) Assuming Molecular Weight Values of 250 for AA, 2000 for BB, and 90 for CC Initial concentration of A 2

c1

?4

3

8083 7.97 4750 5.00 3492 3.88 2862 3.32 2660 3.14

1

3 12 36

5427 5.60 3625 4.00 2926 3.38 2616 3.10 2540 3.03

4

0% Conversion 4505 4.78 3250 3.67 2772 3.24 2571 3.07 2524 3.02

5

6

403 1 4.36 3062 3.50 2699 3.18 2551 3.04 2517 3.02

3 740 4.10 2950 3.40 2657 3.14 2540 3.04 2513 3.01

95% Conversion 36,400 55.9 26,900 41.5 26,400 41.0 26,400 41.0 26,400 41.0

'/3

1 3 12 36

20,600 42.0 15,300 31.2 15,030 30.7 15,030 30.7 15,030 30.7

15,300 37.3 11,400 27.8 11,200 27.3 11,200 27.3 11,200 27.3

12,700 35.0 9,450 26.1 9,260 25.6 9,260 25.6 9,260 25.6

11,110 33.6 8,270 25.1 8,100 24.5 8,100 24.5 8,100 24.5

61,200 169.5 45,600 126.4 45,500 125.9 45,400 125.7 45,400 125.7

53,600 162.7 39,900 121.2 39,700 120.7 39,700 120.7 39,700 120.7

99% Conversion 175,000 270.7 131,000 201.7 130,000 201.2 130,000 201.2 130,000 201.2

Y3

1 3 12 36

99,400 203.2 73,950 151.4 73,700 150.9 73,700 150.9 73,700 150.9

73,900 180.7 55,000 134.6 54,800 134.1 54,800 134.1 54,800 134.1

In order to meet the criterion that the same is obtained when s - 1, s 2 2 is used to count the number of extender molecules per hard block segment instead of s, the following equation must be true.

s =2

reveals that only the most probable distribution conforms to eq 7 and 8. Thus m

s=2

This result immediately implies

m

C(r + 1 - a ) F ( r ) =

p"-'/(l

-

p)

3

m

C(r + 1 - a ) * F ( r ) = pa-1(1 + p)/(l - p )

s-2

s=l

(12)

r=o

and

CsG(s) 9=1

=

KC(S - 1)G(s)

(8)

s=2

where K is a constant. Examination of the distributions in common use in macromolecular chemistry found in ref 12

where a = 1, 2 , 3 , etc. Hence the distribution of extender molecules in the hard block segment must conform to the most probable distribution. The weight average number of extender molecules per hard block segment is then

s,

=

(1 + p)/(l

-

p)

(13)

Sequence Length Distribution in Segmented Block Copolymers 877

Vol. 7, No. 6, November-December 1974

Table VI Number Average Molecular Weight a n d Degree of Polymerization of “Polymer” as a Function of the Extending Reacition for t h e Nonsymmetrical Casea

Table VI11 Number Average Length of Hard Block Segments as a Function of Initial Concentration, Conversion, a n d Reactivity Ratio for the Nonsymmetrical Case, Model I Initial concentration of A,

Initial concentration of A, 2

li.

4

:3

3

4225 334’9 3063 4.53 :3.75 3.50 2717 1 2 3254 283:3 3.67 3.30 3.19 36 2842 26213 2582 3.30 3.11 3.07 Same values of molecular weight of mers as in Table V.

5

6

2922 2837 3.37 3.30 2661 2628 3.14 3.11 2560 2547 3.05 3.04 monomers and prepoly-

Table VI1 Number Average L,ength of H a r d Block Segments as a Function of Initial1 Concentration, Conversion, a n d Reactivity Ratio for the Symmetrical Case, Model I (see also Figure 1)

Initial concentration of A,

1 3 12 36

2.63 1.82 1.40 1.15 1.07

3.21 2.51 2.15 1.96 1.90

3.69 3.10 2.82 2.67 2.63

4.09 3.60 3.39 3.28 3.25

4.45 4.03 3.88 3.80 3.78

99% Conversion ‘/3

1 3 12 36

3.26 1.96 1.43 1.16 1.07

4.22 2.89 2.33 2.07 2.01

5.10 3.77 3.24 3.00 2.94

5.93 4.63 4.12 3.90 3.84

6.73 5.46 4.97 4.76 4.71

6.71 5.00 4.35 4.09 4.03 4.00

7.76 6.00 5.35 5.09 5.03 5.00

100% Conversion ‘/3

1 3 12 36 Ideal

3.48 2.00 1.44 1.16 1.07 1.00

4.60 3.00 2.38 2.10 2.03 2.00

5.60 4.00 3.36 3.10 3.04 3.00

and the heterogeneity index is

s,/s

=:

1

+p

= 1

2

3 12 36

1.67 1.31 1.14

3

4

95% Conversion 2.41 3.04 2.10 2.80 1.97 2.68

5

6

3.57 3.38 3.30

4.04 3.89 3.82

4.45 4.07 3.91

5.30 4.93 4.78

4.75 4.29 4.11 4.00

5.75 5.29 5.11 5.00

99% Conversion 3 12 36

1.75 1.33 1.15

2.68 2.25 2.08

3.58 3.18 3.01

100% Conversion 3 12 36 Ideal

1.77 1.34 1.15 1.00

2.75 2.30 2.11 2.00

3.75 3.29 3.11 3.00

Table IX Heterogeneity Index, &/s, for the Symmetrical Case at 100% Conversion

95% Conversion ‘13

P

+ X?/G

Initial concentration of A, ~

P ‘/3

1 3 12 36

2

3

4

1.71 1.50 1.31 1.14 1.07

1.78 1.67 1.58 1.52 1.51

1.82 1.75 1.70 1.68 1.67

5 1.85 1.80 1.77 1.76 1.75

6 1.87 1.83 1.81 1.80 1.80

Table X Heterogeneity Index for the Nonsymmetrical Case at 100%Conversion Initial concentration of A, li.

2

1 3 12 36

1.50 1.43 1.25 1.13

3 1.69 1.64 1.56 1.53

4 1.75 1.73 1.70 1.68

5 1.80 1.79 1.77 1.76

6 1.83 1.83 1.81 1.80

(14)

Because the parametler p is not close to unity, the ratio of weight to number average sequence length is not equal to 2. Calculated values of the heterogeneity index are given in Tables IX and X for the symmetric and nonsymmetric cases. The distribution is narrower for low values of AI and high values of p, that is, conditions which favor low values of XI? The value of s for .models I1 and I11 for either the symmetrical or the nonsymmetrical case as a function of conversion can be obtained easily. Figure 1 shows the dependence of S as a function of conversion for the symmetrical case with h = 1h or 36 under stoichiometric conditions for the different models. The length of the hard block segment is more difficult to control a t large values of A I and a t small

values of h. Because the various models of the hard block segments give essentially the same values, there is little departure, if any, from the most probable distribution in the range of 95-100% conversion. Similar conclusions can be obtained from an examination of S as a function of conversion for the nonsymmetrical case, model I, Table VIII. Models I1 and I11 differ from model I (Table XI) by even smaller amounts than in the symmetrical case. Examination of Tables VI1 and VI11 a t 100% conversion immediately shows that S is larger than that for the ideal case where p is finite. This means that properties which depend upon the hard block length will show this dependence a t lower values of the A1IB1 ratio than predicted by the ideal case. In the two-stage polymerization considered thus

878 Peebles

Macromolecules Table XI Total Number a n d Internal Number of Hard Block Sequences per Polymer Molecule for the Symmetrical Case, Model I

Initial concentration of A 2

P Total

y3 1 3 12 36

3

Internal

3.82 2.85 2.24 1.24 0.64

2.07 1.93 1.75 1.03 0.54

4

Total

Internal

3.45 2.77 2.76 2.75 2.74

1.63 1.58 1.82 1.93 1.95

Total

6

5

Internal

95% Conversion 3.17 1.32 1.24 2.53 1.42 2.56 1.52 2.60 2.61 1.56

Total

Internal

Total

Internal

2.95 2.33 2.33 2.36 2.37

1.10 0.98 1.11 1.18 1.20

2.78 2.17 2.15 2.17 2.17

0.93 0.80 0.88 0.93 0.94

10.84 8.73 9.42 9.77 9.85

8.55 7.19 8.04 8.44 8.54

8.80 7.71 8.23 8.49 8.55

6.73 6.13 6.77 7.06 7.14

99% Conversion

v3 1 3 12 36

14.81 12.86 10.75 6.02 4.20

12.92 11.84 10.24 5.81 4.07

12.58 11.66 12.52 12.74 12.75

10.60 10.36 11.51 11.86 11.91

A

10.96 10.03 10.92 11.35 11.44

B

@=36

S

7.0

6.0

5.0

4.0

8.93 8.58 9.67 10.18 10.31

mer, Az, with an excess amount of extender, Cz, to form an oligomer of the form CC(AACC)t. This mixture can then be combined with the reaction mixture from the Al-Bl reaction to form the final polymer. This scheme has been incorporated into the computer program to allow one to calculate the number average molecular weight and the number average length of the hard block segments under a variety of polymerization conditions. In general, if s depends on conversion in the two-step synthesis, the dependence is increased in the three-step synthesis. On the other hand, a t low values of total diisocyanate concentration, long sequence lengths should be able to be produced. A.n alternative method of forming a segmented block copolymer is to combine all three reactants simultaneously in a one-stage polymerization. For the most general case one can write .U

2.0

A A -I- CC AACC

t 95

97

99

I .o

'.O

I00

% CONVERSION

I

I

I

95

97

99

*

+ BB AABB + BB A BBAABB

3.0

I1 1

% CONVERSION

Figure 1. The number average number of counted units in a hard block segment, 8, as a function of conversion for the symmetrical case when p = % and p = 36, under stoichiometric conditions, B1 = 1 molh., A 1 = 2, 3, 4, 5, and 6 mol/l. in ascending order. For p = %: j . 1 = j. 11 (--) and 111 (. .). For p = 36: 3 I = 9 111 (--) and 3 11 ( - - -). The method of counting the units varies among cases I, 11, and 111.

.

far, on increasing the ratio A1/BI to obtain longer hard block segments, the molecular weight of the resultant extended polymer is reduced. Can a recipe be found in which the average hard block length is varied while maintaining the molecular weight reasonably constant? One method would be to react varying amounts of diisocyanate mono-

+ CC

*

AACC

CCAACC

That is, the first reaction of the monomer AA with BB in the symmetrical case has a specific rate constant p k , while the rate of AA with CC is y u k . The specific rate constant for a polymer chain with an A-type end reacting with a Btype end is k, while that reacting with a C-type end is uk. The value of u probably is different from unity; see Table I. Appendix C gives the derivation for S for the limiting case when Y = 1. At 100% conversion under stoichiometric conditions -

s = AJB,

0.15)

which is independent of p and equal to the two-stage process when w = 1. Of course, as u departs from unity, deviations from eq 15 will occur. The nonlinear increase in as a function of conversion a t low values of f i and high values of A I shown in Figure 1 is caused by coupling of hard block segment types 5-9 in Table I1 in the final steps of conversion. The total concentration of hard block segments is given by Zg(s)I, while the

Vol. 7,No.6,Nouem.ber-December1974

7

Sequence Length Distribution in Segmented Block Copolymers 879

880 Peebles

Macromolecules

*a-

Vol. 7,No.6,Nouember-December 1974

Sequence Lenght Distribution in Segmented Block Copolymers

881

d::

total concentration of internal segments is given by zO(s) (see eq A62 and A63). Division of Zg(s)' or XO(s) by the concentration of polymer molecules yields the total and internal number of hard block segments per polymer molecule. For A I = 6 molfl., only about 40% of the hard block segments are internal sequences a t 95% conversion, the remaining 60% being terminal chains or isolated sequences. From 95 to 99% conversion, the per cent of internal sequences raises from 4.0 to 80%. Of course, at 100% conversion, all hard block segments are internal.

Conclusions The sequence length distribution for the hard block segment in sequential block copolymers prepared by a twostage process has been calculated from the derived kinetic expressions. Two cases are considered: one in which the diisocyanate monomer has a reactivity different from that of a terminal isocyanate group in a polymer chain, the other in which the reactivities of the isocyanate groups in the monomer differ because of steric reasons. Because there are no major differences in the results of the two cases, it is not considered worthwhile to extend the treatment by combining the two cases. A number of models of the hard blocks are considered wherein the method of counting the repeat unit is varied. All models considered lead to the same number average sequence length at 100% reaction under stoichiometric conditions. Such a result can only occur if the distribution of repeat units in the hard block segment is of the most probable foi-m. The heterogeneity index of the hard block segments is less than 2 because only short se-

quence lengths are produced. Control of the sequence length and the molecular weight of the polymer can be affected by altering the proportions of diisocyanate, prepolymer, and extender and by the method of mixing the reactants. Acknowledgments. I would like to thank Dr. N. S. Schneider, Army Materials and Mechanics Research Center, for suggesting the problem and Dr. Schneider and Mrs. Joyce Illinger for many stimulating discussions during the course of this work, in particular the suggestion that more than one model of the hard block segment should be considered. Professor J. J. Hermans, University of North Carolina, derived eq A21 in an unpublished work (1963) which formed the foundation for the present work. The arguments leading to its derivation in Appendix A are essentially those used by Professor Hermans. Permission to cite that work was granted by the Chemstrand Research Center, Inc. The cost for the computer calculations was supported by the Department of Chemical Engineering, MIT, and is greatly appreciated. I especially want to acknowledge the assistance of those mentioned above and the reviewers for helpful comments concerning the form of the manuscript. Supplementary and Miniprint Material Available. The programs for computing Zg(s) for both the symmetrical and nonsymmetrical cases as a function of conversion will appear immediately following this article in the microfilm edition of this volume of the journal. Photocopies of the supplementary material and full-sized photocopies of the miniprint material from this paper only or microfiche (105 X 148 mm, 24X reduction, negatives) containing all

882 Sharda, Tschoegl

Macromolecules

of the supplementary material for the papers in this issue may be

obtained from the Journals Department, American Chemical Society, 1155 16th St., N.W., Washington, D. C. 20036. Remit check or money order for $6.00 for photocopy or $2.00 for microfiche, referring to code number MACRO-74-872.

References and Notes (1) Visiting Scientist, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Mass. ( 2 ) G. M. Estes, S.L. Cooper, and A. V. Tobolsky, J. Macromol. Sei., Reu.

Macromol C h e m , 4,313 (1970). ( 3 ) (a) J. M. Buist and H. Gadgeon, “Advances in Polyurethane Technology,” Wiley, New York, N. Y., 1968 (b) L. L. Harrell, Jr., Macromole-

cules, 2,607 (1969).

(4) (a) A. M. North and J. C. Reid, Eur. Polym. J., 8, 1129 (1972); (b) J . Ferguson and D. Patsavoudis, ibid., 8, 385 (1972); (c) C. H. Hespe, E. Meisert, U. Eisele, L. Morbitzer, and W. Goyert. Kolloid-2. Z. Poljm. 250,797 (1972). ( 5 ) R. W. Lenz, “Organic Chemistry of Synthetic High Polymers,” Interscience, New York, N. Y., 1967. (6) L. L. Ferstandig and R. A. Scherrer, J. Amer. Chem. Soc., 81, 4838 (1959). (7) F. H. Brock, J . Org. Chem., 24, 1802 (1959). (8) L. C. Case, J . Polym. Sei., 29,455 (1958). (9) L. C. Case, J.Polym. Sci., 37,147 (1959). (10) P. J. Flory, J . Amer. Chem. Soc., 58,1877 (1936). (11) Anon., 1130 Scientific Subroutine Package, Programmer’s Manual, copyright International Business Machines Corp. (12) L. H. Peebles, Jr., “Molecular Weight Distribution in Polymers,” Inter-

science, New York, N. Y., 1971,.

Thermodynamic Potential Functions for Elastomers S. C. Sharda and N. W. Tschoegl* Department of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91109. Received May 23, 1974

ABSTRACT: The thermodynamic description of the large principal deformations of elastomers requires four thermodynamic energy functions and their associated free energies. The significance of, and relationship between, these potential functions is discussed and their interrelations are derived. The internal energy contribution to the retractive force or the extension in an elastomer is used as an example of the application of the concepts introduced.

T h e thermodynamic (more precisely thermostatic) description of the deformation of a solid system is more complex than that of a fluid (gas or liquid) system. In the latter, volume and pressure are the only mechanical parameters to be taken into account. In the former, the set of mechanical parameters must be enlarged, in the general case, to the 6 6 = 12 components of the strain and stress tens0rs.I This paper discusses the thermodynamic potential functions which are useful in describing the thermodynamics of large principal deformations of elastomers and the general relations between their differentials and those of various sets of independent variables. Because a n elastomer may be regarded as a homogeneous isotropic solid, its deformation can be described more simply than that of a general solid. The potential functions for the thermodynamics of elastomers introduced here are defined in a way which makes them consistent with the definitions of the potential functions in the thermodynamics of fluids. The following discussion aims to provide an exhaustive treatment of the formalism of the thermodynamics of elastomers.

where T i s the (absolute) temperature and S is the entropy. Flory2 has shown that the element of elastic work, dW, done by a homogeneous isotropic system in a large principal deformation is given by 3

d W = - V x f i d In h i

+

Potential Functions The first law of thermodynamics gives the change in the internal energy U as dC‘ = dQ - d W (1) where dQ is the element of heat absorbed, and d W is the element of work done, by the system on its surroundings. If the process is conducted reversibly dQ = TdS (2 1

i=l

(3)

where V is the deformed volume, the t , are the principal true stresses, and the A, are the principal extension ratios defined as the ratios of the stretched lengths, L,, to the unstretched lengths, L,o. Thus, the change in the internal energy is given by d U = TdS

+

V C t i d In X i i

and the change in the associated (Helmholtz) free energy, defined by

A = LT

-

TS

(5)

becomes

dA = - S d T

+

V x t i d In h i i

The last terms in eq 4 and 6 contain both the work of extension and the work of expansion. The two must be separated because changes in volume may be induced both by the application of forces or extensions and by changes in temperature. We consider that the principal true stresses are

ti = ( L i f i / V ) - P

(7)